Plotting in R

Explore different types of plots in ggplot2

ggplot2 is an R package for creating graphics based on The Grammar of Graphics¹. The Grammar of Graphics is a language for talking about the different parts of a plot, and allow you to build plots creatively and iterively.

Required packages

Make sure that you have the following packages installed:

palmerpenguins

DT

tidyverse

ggpubr

plotly

dygraphs

Dataset

We will use the palmerpenguins dataset. Data were collected and made available by Dr. Kristen Gorman and the Palmer Station, Antarctica LTER, a member of the Long Term Ecological Research Network. Alison Horst gathered the data into an R package and is responsible for all the great penguin illustrations.

We will briefly check the structure of the data table (penguins) before we start plotting. Here I am using the datatable function from from DT package which facilitates the display of dataframes, matrices or tibbles on HTLM pages.

library(palmerpenguins)  # load palmerpenguins package
library(DT)  # load DT package
datatable(penguins)  #check table structure

And summarize the penguins table using the summary function:

summary(penguins)  #summarize data

      species          island    bill_length_mm  bill_depth_mm  
 Adelie   :152   Biscoe   :168   Min.   :32.10   Min.   :13.10  
 Chinstrap: 68   Dream    :124   1st Qu.:39.23   1st Qu.:15.60  
 Gentoo   :124   Torgersen: 52   Median :44.45   Median :17.30  
                                 Mean   :43.92   Mean   :17.15  
                                 3rd Qu.:48.50   3rd Qu.:18.70  
                                 Max.   :59.60   Max.   :21.50  
                                 NA's   :2       NA's   :2      
 flipper_length_mm  body_mass_g       sex           year     
 Min.   :172.0     Min.   :2700   female:165   Min.   :2007  
 1st Qu.:190.0     1st Qu.:3550   male  :168   1st Qu.:2007  
 Median :197.0     Median :4050   NA's  : 11   Median :2008  
 Mean   :200.9     Mean   :4202                Mean   :2008  
 3rd Qu.:213.0     3rd Qu.:4750                3rd Qu.:2009  
 Max.   :231.0     Max.   :6300                Max.   :2009  
 NA's   :2         NA's   :2

As we can see from the summary table three different species of penguins were recorded in three different islands.

Scatterplot

Let’s explore if there is a correlation between the body mass of the penguins and the flipper length of the penguins:

library(tidyverse)  # load the tidyverse package; contains ggplot2

ggplot(penguins, aes(x = flipper_length_mm, y = body_mass_g)) + geom_point()

Let’s add the trend line (fitting linear model):

ggplot(penguins, aes(x = flipper_length_mm, y = body_mass_g)) + geom_point() + geom_smooth(method = "lm")

Let’s add a trend line together with the equation and the R² value:

library(ggpubr)  #package the facilitates the display of the equation

ggplot(penguins, aes(x = flipper_length_mm, y = body_mass_g)) + geom_point() + geom_smooth(method = "lm") +
    stat_regline_equation(label.y = 6000, aes(label = ..eq.label..)) + stat_regline_equation(label.y = 5600,
    aes(label = ..rr.label..))

Are there any differences between the species?

# regression equations will overlap, we will use faceting for them (below)
ggplot(penguins, aes(x = flipper_length_mm, y = body_mass_g, color = species)) +
    geom_point() + geom_smooth(method = "lm")

Besides of using different colors for the data points, we can also use different shapes:

ggplot(penguins, aes(x = flipper_length_mm, y = body_mass_g, color = species, shape = sex)) +
    geom_point()

Geoms that draw points have a shape parameter. Legal shape values are the numbers 0 to 25, and the numbers 32 to 127.

Shapes 0 to 14 are outline only: use color to change colors (outline)
Shapes 15 to 20 are fill only: use color to change colors (fill)
Shapes 21 to 25 are outline + fill: use color to change the outline color and fill to change the fill color

Shapes 32 to 127 represent the corresponding ASCII characters: We can also change the point size:

ggplot(penguins, aes(x = flipper_length_mm, y = bill_length_mm, color = species)) +
    geom_point(aes(shape = sex, size = body_mass_g))

The Grammar of Graphics

In the exploration of the palmerpenguins data, we started with a simple plot and added to it. We added a linear model as a trend line, added model parameters to the plot, grouped the data by species, and changed things like point size and shape.

Starting with the first plot, we started by using the ggplot() function to create a ggplot object. As parameters, we told ggplot() we wanted to use penguins as data for the plot, and used the aes() function to define how we wanted to map the penguin data to the plot aesthetics. We mapped flipper_length_mm onto the x axis and body_mass_g onto the y axis.

Next, we have to tell ggplot how we want to display the data. Geoms take mapped data and make it visible on the plot. geom_point() is a geom that (you guessed it) plots points. Note that we’ve added our first layer to the plot by sending the object created by ggplot() to geom_point() using the +. Why not use the pipe (%>%)? Ggplot was developed before the magrittr pipe, so + it is. This has made a lot of people very angry and been widely regarded as a bad move.

Layers are functions, so they take parameters that control what they do. For instance, when we used geom_point() as a layer to display available plot symbols above, we used this line:

geom_point(aes(shape = shape), size = 5, fill = "red")

This uses aes() to map shape from the data to the shape displayed. size = 5 and fill = 'red' define the size and fill color of all points plotted. Assigning a constant to an aesthetic sets it for the entire geom, while mapping data to an aesthetic with aes() allows it to vary with the data mapped.

Each additional layer adds something to the plot or modifies the plot defaults. We can add layers to add additional plots with the same or different data mapped to the plot aesthetics, modify the plot scales, change the coordinate system of the axes, change the theme of the plot, and break the plot into subplots by a categorical variable, which we’ll look at next.

Faceting

Faceting is the process that split the chart window in several small parts (a grid), and display a similar chart in each section. Each section usually shows the same graph for a specific group of the dataset.

# regression equations will overlap, we will use faceting for them
ggplot(penguins, aes(x = flipper_length_mm, y = body_mass_g, color = species)) +
    geom_point(aes(shape = sex)) + facet_wrap(~species, scales = "free_x") + geom_smooth(method = "lm",
    se = FALSE) + stat_regline_equation(label.y = 6000, aes(label = ..eq.label..)) +
    stat_regline_equation(label.y = 5800, aes(label = ..rr.label..))

# regression equations will overlap, we will use faceting for them
ggplot(penguins, aes(x = flipper_length_mm, y = body_mass_g, color = species)) +
    geom_point(aes(shape = sex)) + facet_grid(sex ~ species, scales = "free_x") +
    geom_smooth(method = "lm", se = FALSE)

Themes

There are built in-ggplot themes, as well as theme packages. There is a long list of cosmetic changes you can make with theme(). Let’s try changing themes in other type of plot, histograms using geom_histogram(). Let’s plot the distribution of the flipper length for each species. We will use my favourite theme, them_bw():

ggplot(penguins, aes(flipper_length_mm, fill = species)) + geom_histogram(alpha = 0.6,
    position = "identity") + theme_bw()

theme_void() is another build-in package:

ggplot(penguins, aes(flipper_length_mm, fill = species)) + geom_histogram(alpha = 0.6,
    position = "identity") + theme_void()

As we will see below, themes can be modified.

Labels

Labels can be modified using labs(x = "Title on x axis", y = "Title on y axis")and theme(axis.title.x = element_text(family, face, colour, size), axis.title.y = element_text(family, face, colour, size)):

ggplot(penguins, aes(flipper_length_mm, fill = species)) + geom_histogram(alpha = 0.6,
    position = "identity") + theme_bw() + labs(x = "Flipper length (mm)", y = "Counts") +
    theme(axis.title.x = element_text(color = "grey30", face = "bold", size = 14),
        axis.title.y = element_text(color = "grey30", face = "bold", size = 14))

We can aslo can change the angle, and justification of the axis labels:

# modify axis font size
ggplot(penguins, aes(flipper_length_mm, fill = species)) + geom_histogram(alpha = 0.6,
    position = "identity") + theme_bw() + labs(x = "Flipper length (mm)", y = "Counts") +
    theme(axis.text.x = element_text(color = "black", size = 12, angle = 45), axis.text.y = element_text(color = "black",
        size = 12))

The horizontal or vertical justification, (hjust and vjust) can also be adjusted. This hjust and vjust argument can be best explained using this figure [Source from Stackoverflow]:

e.g.:

ggplot(penguins, aes(flipper_length_mm, fill = species)) + geom_histogram(alpha = 0.6,
    position = "identity") + theme_bw() + labs(x = "Flipper length (mm)", y = "Counts") +
    theme(axis.text.x = element_text(color = "black", size = 12, angle = 45, hjust = 1,
        vjust = 1), axis.text.y = element_text(color = "black", size = 12), axis.title.x = element_text(color = "grey30",
        face = "bold", size = 14), axis.title.y = element_text(color = "grey30",
        face = "bold", size = 14))

Axis

The appearance of the text on the axis can be modified using theme(axis.text.x = element_text(family, face, colour, size), axis.text.y = element_text(family, face, colour, size)):

# modify axis font size
ggplot(penguins, aes(flipper_length_mm, fill = species)) + geom_histogram(alpha = 0.6,
    position = "identity") + theme_bw() + labs(x = "Flipper length (mm)", y = "Counts") +
    theme(axis.text.x = element_text(color = "black", size = 12), axis.text.y = element_text(color = "black",
        size = 12), axis.title.x = element_text(color = "black", face = "bold", size = 14),
        axis.title.y = element_text(color = "black", face = "bold", size = 14))

Legends

Legends can be modified inside theme using legend.title=element_text(family, face, size, color) and legend.text=element_text(family, face, size, color). The position of the legend can be modified using legend.position:

ggplot(penguins, aes(flipper_length_mm, fill = species)) + geom_histogram(alpha = 0.6,
    position = "identity") + theme_bw() + labs(x = "Flipper length (mm)", y = "Counts",
    fill = "Species") + theme(axis.text.x = element_text(color = "black", size = 12),
    axis.text.y = element_text(color = "black", size = 12), axis.title.x = element_text(color = "black",
        face = "bold", size = 14), axis.title.y = element_text(color = "black", face = "bold",
        size = 14), legend.title = element_text(color = "black", face = "bold", size = 14),
    legend.text = element_text(size = 12), legend.position = "top")

It is also possible to position the legend inside the plotting area.The numeric position below is relative to the entire area, including titles and labels, not just the plotting area; where x,y is 0,0 (bottom left) to 1,1 (top right):

ggplot(penguins, aes(flipper_length_mm, fill = species)) + geom_histogram(alpha = 0.6,
    position = "identity") + theme_bw() + labs(x = "Flipper length (mm)", y = "Counts",
    fill = "Species") + theme(axis.text.x = element_text(color = "black", size = 12),
    axis.text.y = element_text(color = "black", size = 12), axis.title.x = element_text(color = "black",
        face = "bold", size = 14), axis.title.y = element_text(color = "black", face = "bold",
        size = 14), legend.title = element_text(color = "black", face = "bold", size = 14),
    legend.text = element_text(size = 12), legend.position = c(0.9, 0.85))

Colors

There are several ways colors can be modified in ggplot2.

Manually with scale_color_manual() or scale_fill_manual() (can accept hex numbers or names):

ggplot(penguins, aes(flipper_length_mm, fill = species)) + geom_histogram(alpha = 0.6,
    position = "identity") + scale_fill_manual(values = c("#69b3a2", "purple", "orange")) +
    theme_bw() + labs(x = "Flipper length (mm)", y = "Counts", fill = "Species") +
    theme(axis.text.x = element_text(color = "black", size = 12), axis.text.y = element_text(color = "black",
        size = 12), axis.title.x = element_text(color = "black", face = "bold", size = 14),
        axis.title.y = element_text(color = "black", face = "bold", size = 14), legend.title = element_text(color = "black",
            face = "bold", size = 14), legend.text = element_text(size = 12), legend.position = c(0.9,
            0.85))

Creating of evenly spaced colours for discrete data with scale_color_hue() or scale_fill_hue() (can accept hex or names):

ggplot(penguins, aes(flipper_length_mm, fill = species)) + geom_histogram(alpha = 0.6,
    position = "identity") + scale_fill_hue(h = c(0, 90)) + theme_bw() + labs(x = "Flipper length (mm)",
    y = "Counts", fill = "Species") + theme(axis.text.x = element_text(color = "black",
    size = 12), axis.text.y = element_text(color = "black", size = 12), axis.title.x = element_text(color = "black",
    face = "bold", size = 14), axis.title.y = element_text(color = "black", face = "bold",
    size = 14), legend.title = element_text(color = "black", face = "bold", size = 14),
    legend.text = element_text(size = 12), legend.position = c(0.9, 0.85))

Using packages that contain palettes e.g. RColorBrewer, viridis, or Paletteer:

library(viridis)
ggplot(penguins, aes(flipper_length_mm, fill = species)) + geom_histogram(alpha = 0.6,
    position = "identity") + scale_fill_viridis(discrete = TRUE) + theme_bw() + labs(x = "Flipper length (mm)",
    y = "Counts", fill = "Species") + theme(axis.text.x = element_text(color = "black",
    size = 12), axis.text.y = element_text(color = "black", size = 12), axis.title.x = element_text(color = "black",
    face = "bold", size = 14), axis.title.y = element_text(color = "black", face = "bold",
    size = 14), legend.title = element_text(color = "black", face = "bold", size = 14),
    legend.text = element_text(size = 12), legend.position = c(0.9, 0.85))

Barplot

A simple barplot can be created using the function geom_bar(). We will plot how many individuals from each species were recorded in each island:

ggplot(penguins, aes(x = island, fill = species)) + geom_bar() + scale_fill_manual(values = c("#69b3a2",
    "purple", "orange")) + theme_bw() + labs(x = "Island", y = "Number of individuals") +
    theme(axis.text.x = element_text(color = "black", size = 12), axis.text.y = element_text(color = "black",
        size = 12), axis.title.x = element_text(color = "black", face = "bold", size = 14),
        axis.title.y = element_text(color = "black", face = "bold", size = 14))

If we want to produce an non-stacked barplot we need to use the argument position=position_dodge2 in geom_bar():

ggplot(penguins, aes(x = island, fill = species)) + geom_bar(position = position_dodge2(preserve = "single")) +
    scale_fill_manual(values = c("#69b3a2", "purple", "orange")) + theme_bw() + labs(x = "Island",
    y = "Number of individuals") + theme(axis.text.x = element_text(color = "black",
    size = 12), axis.text.y = element_text(color = "black", size = 12), axis.title.x = element_text(color = "black",
    face = "bold", size = 14), axis.title.y = element_text(color = "black", face = "bold",
    size = 14))

Changing the position to fill, will give us the relative abundance in a stacked plot:

ggplot(penguins, aes(x = island, fill = species)) + geom_bar(position = "fill") +
    scale_fill_manual(values = c("#69b3a2", "purple", "orange")) + theme_bw() + labs(x = "Island",
    y = "Relative abundance") + theme(axis.text.x = element_text(color = "black",
    size = 12), axis.text.y = element_text(color = "black", size = 12), axis.title.x = element_text(color = "black",
    face = "bold", size = 14), axis.title.y = element_text(color = "black", face = "bold",
    size = 14))

We can aslo create the same plot horizontally with coord_flip():

ggplot(penguins, aes(x = island, fill = species)) + geom_bar(position = "fill") +
    scale_fill_manual(values = c("#69b3a2", "purple", "orange")) + theme_bw() + labs(x = "Island",
    y = "Relative abundance") + theme(axis.text.x = element_text(color = "black",
    size = 12), axis.text.y = element_text(color = "black", size = 12), axis.title.x = element_text(color = "black",
    face = "bold", size = 14), axis.title.y = element_text(color = "black", face = "bold",
    size = 14)) + coord_flip()

Boxplots

Let visualize the distribution of flipper length in the different species and sexes using the function geom_boxplot() to create a box and whiskers plot:

ggplot(penguins, aes(x = species, y = flipper_length_mm, fill = sex)) + geom_boxplot() +
    theme_bw() + labs(x = "Species", y = "Flipper length (mm)") + theme(axis.text.x = element_text(color = "black",
    size = 12), axis.text.y = element_text(color = "black", size = 12), axis.title.x = element_text(color = "black",
    face = "bold", size = 14), axis.title.y = element_text(color = "black", face = "bold",
    size = 14))

Let’s omit the NA values during plotting and reduce the outlier size:

ggplot(na.omit(penguins), aes(x = species, y = flipper_length_mm, fill = sex)) +
    geom_boxplot(outlier.size = 1) + theme_bw() + labs(x = "Species", y = "Flipper length (mm)") +
    theme(axis.text.x = element_text(color = "black", size = 12), axis.text.y = element_text(color = "black",
        size = 12), axis.title.x = element_text(color = "black", face = "bold", size = 14),
        axis.title.y = element_text(color = "black", face = "bold", size = 14))

ggplot(na.omit(penguins), aes(x = species, y = flipper_length_mm, fill = sex)) +
    geom_boxplot() + geom_jitter(color = "black", size = 0.4, alpha = 0.9) + theme_bw() +
    labs(x = "Species", y = "Flipper length (mm)") + theme(axis.text.x = element_text(color = "black",
    size = 12), axis.text.y = element_text(color = "black", size = 12), axis.title.x = element_text(color = "black",
    face = "bold", size = 14), axis.title.y = element_text(color = "black", face = "bold",
    size = 14))

Violin plots

We can visualize the same distribution using the function geom_violin() to create a violin plot, a mirrored density plot displayed in the same way as a boxplot:

ggplot(na.omit(penguins), aes(x = species, y = flipper_length_mm, fill = sex)) +
    geom_violin() + theme_bw() + labs(x = "Species", y = "Flipper length (mm)") +
    theme(axis.text.x = element_text(color = "black", size = 12), axis.text.y = element_text(color = "black",
        size = 12), axis.title.x = element_text(color = "black", face = "bold", size = 14),
        axis.title.y = element_text(color = "black", face = "bold", size = 14))

We also overlay different types of plots:

ggplot(na.omit(penguins), aes(x = species, y = flipper_length_mm, fill = sex)) +
    geom_violin() + geom_boxplot(position = position_dodge(width = 0.9), width = 0.2) +
    theme_bw() + labs(x = "Species", y = "Flipper length (mm)") + theme(axis.text.x = element_text(color = "black",
    size = 12), axis.text.y = element_text(color = "black", size = 12), axis.title.x = element_text(color = "black",
    face = "bold", size = 14), axis.title.y = element_text(color = "black", face = "bold",
    size = 14))

Plot interactively with plotly or dygraphs

##Dataset The following material is from XXX

Get URL to CSV

Visit the ERDDAP server https://oceanview.pfeg.noaa.gov/erddap and do a Full Text Search for Datasets using “cciea” in the text box before clicking Search. These are all the California Current IEA datasets. From the listing of datasets, click on data for the “CCIEA Anthropogenic Drivers” dataset. Note the filtering options for time and other variables like consumption_fish (Millions of metric tons) and cps_landings_coastwide (1000s metric tons). Set the time filter from being only the most recent time to the entire range of available time. Scroll to the bottom and Submit with the default .htmlTable view. You get an web table of the data. Notice the many missing values in earlier years. Go back in your browser to change the the File type to .csv. Now instead of clicking Submit, click on Just generate the URL. Although the generated URL lists all variables to include, the default is to do that, so we can strip off everything after the .csv, starting with the query parameters ? .

Download CSV

Let’s use this URL to download a new file

# set variables
csv_url <- "https://oceanview.pfeg.noaa.gov/erddap/tabledap/cciea_AC.csv"
# if ERDDAP server down (Error in download.file) with URL above, use this:
# csv_url <-
# 'https://raw.githubusercontent.com/microbiaki/workshop_t2/main/data/cciea_AC.csv'
dir_data <- "data"
# derived variables
csv <- file.path(dir_data, basename(csv_url))
# create directory
dir.create(dir_data, showWarnings = F)
# download file
if (!file.exists(csv)) download.file(csv_url, csv)

Read table

Now open the file by going into the Files RStudio pane, More -> Show Folder in New Window. Then double click on data/cciea_AC.csv to open in your Spreadsheet program (like Microsoft Excel or Apple Pages or LibreOffice Calc).

# attempt to read csv
d <- read.csv(csv)
# show the data frame
head(d)

Note how the presence of the 2nd line with units makes the values character <chr> data type. But we want numeric values. So we could manually delete that second line of units or look at the help documentation for this function (?read.csv in Console pane; or F1 key with cursor on the function in the code editor pane). Notice the skip argument, which we can implement like so:

# read csv by skipping first two lines, so no header
d <- read.csv(csv, skip = 2, header = FALSE)

# update data frame to original column names
names(d) <- names(read.csv(csv))

# fix year
d$time <- sub("-.+", "", d$time)

d$time <- as.integer(d$time)
# update for future reuse (NEW!)
write.csv(d, csv, row.names = F)

datatable(d)

Series line plot

Next, let’s also show the other regional values (CA, OR and WA; not coastwide) in the plot as a series with different colors aes(color = region). To do this, we’ll want to tidy the data into long format so we can have a column for total_fisheries_revenue and another region column to supply as the group and color aesthetics based on aesthetics we see are available for geom_line():

d_rgn <- d %>% 
  # select columns
  select(
    time, 
    starts_with("total_fisheries_revenue")) %>% 
  # exclude column
  select(-total_fisheries_revenue_coastwide) %>% 
  # pivot longer
  pivot_longer(-time) %>% 
  # mutate region by stripping other
  mutate(
    region = name %>% 
      str_replace("total_fisheries_revenue_", "") %>% 
      str_to_upper()) %>% 
  # filter for not NA
  filter(!is.na(value)) %>% 
  # select columns
  select(time, region, value)
  
# create plot object
p_rgn <- ggplot(
  d_rgn,
  # aesthetics
  aes(x= time, y = value, group = region, color = region))+
    theme_bw()+
    labs(x = "Year", y = "Millions $")+
    theme(axis.text.x = element_text(color = "black", size = 12, hjust = 1, vjust = 1),
        axis.text.y = element_text(color = "black", size = 12),
        axis.title.x = element_text(color = "grey30", face = "bold", size = 14),
        axis.title.y = element_text(color = "grey30", face = "bold", size = 14))+
  # geometry
  geom_line()
# show plot
p_rgn

When rendering to HTML, you can render most ggplot objects interactively with ggplotly(). The plotly library is an R htmlwidget providing simple R functions to render interactive JavaScript visualizations.

library(plotly)
ggplotly(p_rgn)

Create interactive time series with `dygraph()`

Another htmlwidget plotting library written more specifically for time series data is dygraphs. Unlike the ggplot2 data input, a series is expected in wide (not tidy long) format. So we use tidyr’s pivot_wider() first.

library(dygraphs)
d_rgn_wide <- d_rgn %>%
    mutate(Year = time) %>%
    select(Year, region, value) %>%
    pivot_wider(names_from = region, values_from = value)
datatable(d_rgn_wide)

d_rgn_wide %>%
    dygraph() %>%
    dyRangeSelector()

Leland Wilkinson. The Grammar of Graphics (Statistics and Computing) 2nd Edition↩︎